Temporal Video Segmentation Using Unsupervised

This paper proposes a content-based temporal video segmentation system that integrates syntactic (domain-independent) and semantic (domain-dependent) features for automatic management of video data. Temporal video segmentation includes scene change detection and shot classiication. The proposed scene change detection method consists of two steps: detection and tracking of semantic objects of interest speciied by the user, and an unsupervised method for detection of cuts, and edit eeects. Object detection and tracking is achieved using a region matching scheme, where the region of interest is deened by the boundary of the object. A new unsupervised scene change detection method based on 2-class clustering is introduced to eliminate the data dependency of threshold selection. The proposed shot classiication approach relies on semantic image features and exploits domain-dependent visual properties such as shape, color, and spatial connguration of tracked semantic objects. The system has been applied to segmentation and classiication of TV programs collected from diier-ent channels. Although the paper focuses on news programs, the method can easily be applied to other TV programs with distinct semantic structure.